Transcription of Training data-efficient image transformers & distillation ...
{{id}} {{{paragraph}}}
Example: biology
Training data-efficient image transformers & distillation ...
Training data-efficient image transformers & distillation through attentionHugo Touvron?, Matthieu Cord Matthijs Douze?Francisco Massa?Alexandre Sablayrolles?Herv e J egou??Facebook AI Sorbonne UniversityAbstractRecently, neural networks purely based on attention were shown to ad-dress image understanding tasks such as image classification. These high-performing vision transformers are pre-trained with hundreds of millionsof images using a large infrastructure, thereby limiting their this work, we produce competitive convolution-free transformers bytraining on Imagenet only. We train them on a single computer in less than3 days. Our reference vision transformer (86M parameters) achieves top-1accuracy of (single-crop) on ImageNet with no external importantly, we introduce a teacher-student strategy specific totransformers.
transformer student by either a convnet or a transformer teacher. We introduce a new distillation procedure specific to transformers and show its superiority. 3 Vision transformer: overview In this section, we briefly recall preliminaries associated with the vision trans-former [15, 52], and further discuss positional encoding and resolution.
Loading..
Tags:
Information
Domain:
Source:
Link to this page:
Please notify us if you found a problem with this document: